La Scimmia Yoga is an Italian online yoga platform created to guide people on a journey of movement, mindfulness, and self-discovery. With classes of varying lengths, styles, and levels, it supports practitioners in building a consistent practice that fits their lifestyle and personal goals.
The Challenge
After subscribing, many users found it difficult to choose the right follow-up course. The breadth of the catalogue, normally a strength, created friction: users felt overwhelmed scrolling through dozens of options, unsure which programme matched their level, objectives, or preferences.
Behind the scenes, the team struggled with an equally complex challenge. Matching each user’s unique needs to the right course used to be a manual and unscalable process. With demand increasing, the team needed an automated, intelligent and privacy-safe way to recommend the perfect courses to every user, without increasing operational workload.
The Solution
To remove friction and boost engagement, Happy Horizon implemented a smart automated recommendation engine, powered by profile data, conversational input, and AI.
New subscribers received a guided “Experience & Objectives” flow, asking three simple but essential questions about:
If a user hadn’t completed their profile, Deployteq automatically triggered a follow-up email encouraging them to fill in the missing details, improving both data quality and recommendation accuracy.
When the answers were submitted in the Deployteq form, they were stored in the Deployteq database and used to generate a personalised prompt for the AI engine. From the full course feed of 75+ lessons, AI matched and returned the top three best-fit courses, which were instantly delivered to the user by email. What previously required manual browsing, or manual selection, now happened in milliseconds, fully automated.
How Deployteq Made It Possible
1. Dynamic profile enrichment
2. Automated webhook integration
User responses were passed into Deployteq via webhook, allowing real-time personalisation. Deployteq then generated a structured prompt containing the user’s values, ready for AI processing—without sending any personal data to the model, maintaining full privacy compliance.
3. AI-powered matching at scale
- A shortlist of the top 3 matching courses
- Course IDs and metadata needed for personalised email delivery
4. Personalised email delivery
The Results
Course recommendation emails achieved a 75.6% open rate, significantly outperforming the already strong 59.5% open rate of the profile-completion emails.
Click performance was equally strong, with a 27.8% CTR and 36.8% CTOR, indicating that recipients actively interacted with the recommended content.
641 completed profiles, resulting in a 19.6% conversion rate from the profile-completion flow. Nearly 1 in 5 new subscribers voluntarily shared preference data, crucial for long-term relevance and CLV growth.
90.3% of all feedback on course recommendation emails was positive.
Extremely low unsubscribe rates (0.09–0.37%)
Feedback from the course recommendation email
“Thank you for the recommendations. I think they suit me. I also hope to be able to follow the Pilates course once it becomes available. Many thanks!”
